WO2015180252A1 - 一种多目标逆向违规检测方法、装置及系统 - Google Patents

一种多目标逆向违规检测方法、装置及系统 Download PDF

Info

Publication number
WO2015180252A1
WO2015180252A1 PCT/CN2014/082767 CN2014082767W WO2015180252A1 WO 2015180252 A1 WO2015180252 A1 WO 2015180252A1 CN 2014082767 W CN2014082767 W CN 2014082767W WO 2015180252 A1 WO2015180252 A1 WO 2015180252A1
Authority
WO
WIPO (PCT)
Prior art keywords
detected
dynamic target
determining
target
time period
Prior art date
Application number
PCT/CN2014/082767
Other languages
English (en)
French (fr)
Inventor
熊培杰
曾刚
熊壮
Original Assignee
中兴通讯股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2015180252A1 publication Critical patent/WO2015180252A1/zh

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/20Analysis of motion

Definitions

  • the present invention relates to the field of video recognition, and in particular, to a multi-target reverse violation detection method, apparatus, and system.
  • BACKGROUND OF THE INVENTION Currently, video surveillance is used in various occasions, such as shopping malls, subways, passenger station entrances and exits, traffic lanes, etc., to monitor events occurring in various occasions.
  • the main provision is An afterthought, the video retreat method to deal with some emergencies, to see if there are violations in the video recording, which not only wastes manpower, consumes time, and also misses some passers-by or vehicles with irregularities.
  • Embodiments of the present invention provide a multi-target reverse violation detection method, apparatus, and system, to at least solve the above technical problem, and can automatically determine whether there is a reverse violation behavior in a video recording according to a monitoring direction set by a user, thereby There is no need for the user to view the behavior of the reverse violation in the video recording, saving manpower.
  • an aspect of the present invention provides a multi-target reverse violation detection method, including: acquiring a video image to be detected; receiving a user operation instruction, determining a to-be-detected area in the to-be-detected video image and a reference moving direction corresponding to each to-be-detected area; determining, for each to-be-detected area, a dynamic target in the to-be-detected area; and determining, for each dynamic target, a moving direction of the dynamic target in a preset time period And determining, according to the reference moving direction corresponding to the to-be-detected area to which the dynamic target belongs, whether the moving direction of the dynamic target in the preset time period is in violation.
  • the multi-objective reverse violation detection method wherein the determining the dynamic target in the to-be-detected area is: acquiring the foreground data in the to-be-detected area by using an inter-frame difference multiplication method; The closing/processing operation processes the foreground data; and performs the segmentation of the connected region on the processed foreground data to obtain a dynamic target in the to-be-detected region.
  • the multi-target reverse violation detection method wherein the to-be-detected video image is a binary image processed by binarization.
  • the multi-target reverse violation detecting method wherein the determining a moving direction of the dynamic target in a preset time period is specifically:: for each frame image of the to-be-detected video image in the preset time period, Determine The dynamic target position; determining a moving direction of the dynamic target within the preset time period by direction fitting according to a position of the dynamic target in each frame image.
  • the multi-objective reverse violation detecting method wherein the determining whether the moving direction of the dynamic target in the preset time period is irregular is: determining an angle between the moving direction and the reference moving direction; When the angle of the angle exceeds the preset angle range, the dynamic target reverse violation is determined.
  • the multi-target reverse violation detection method wherein the determining whether the moving direction of the dynamic target is in violation of a preset time period further comprises: when the moving target violates a movement direction within a preset time period, When the dynamic target is violated, a snapshot of the dynamic target is captured and stored.
  • the multi-target reverse violation detecting apparatus wherein the second determining module includes: an acquiring unit configured to acquire foreground data in the to-be-detected area by using an interframe difference multiplication method; and a preprocessing unit configured to pass the image form The open/close operation in the process processes the foreground data; the area dividing unit is configured to perform the connected area segmentation on the processed foreground data to obtain a dynamic target in the area to be detected.
  • the multi-target reverse violation detecting apparatus wherein the video image to be detected is a binary image subjected to binarization processing.
  • the multi-target reverse violation detecting apparatus wherein the third determining module includes: a first determining unit, configured to determine the dynamic for each frame image of the to-be-detected video image in the preset time period a second determining unit, configured to determine a moving direction of the dynamic target within the preset time period by direction fitting according to a position of the dynamic target in each frame image.
  • the multi-objective reverse violation detecting device wherein the determining module comprises: a third determining unit configured to determine an angle between the moving direction and the reference moving direction; and a determining unit configured to be the angle When the preset angle range is exceeded, the dynamic target reverse violation is determined.
  • the multi-target reverse violation detecting device further comprising: a photographing module, configured to capture a snapshot of the dynamic target when the dynamic target violates the rule when the dynamic target violates a movement direction within a preset time period And store it.
  • a photographing module configured to capture a snapshot of the dynamic target when the dynamic target violates the rule when the dynamic target violates a movement direction within a preset time period And store it.
  • Another aspect of the embodiments of the present invention further provides a multi-target reverse violation detection system, including the multi-target reverse violation detection apparatus and a plurality of cameras, wherein the plurality of cameras are configured to generate a video image, and according to the multi-target The snapshot of the reverse violation detection device is taken.
  • the embodiment of the present invention provides a multi-target reverse violation detection method, device, and system, which can automatically determine whether there is a reverse violation in a video recording according to a monitoring direction set by a user.
  • the behavior thus, eliminates the need for the user to view the behavior of the reverse violation in the video recording, saving manpower.
  • FIG. 1 is a schematic flow chart of a multi-target reverse violation detection method according to Embodiment 1 of the present invention.
  • FIG. 2 is a schematic flowchart diagram of a multi-target reverse violation detection method according to Embodiment 2 of the present invention.
  • FIG. 1 is a schematic flow chart of a multi-target reverse violation detection method according to Embodiment 1 of the present invention.
  • FIG. 2 is a schematic flowchart diagram of a multi-target reverse violation detection method according to Embodiment 2 of the present invention.
  • Embodiment 3 is a schematic structural diagram of a multi-target reverse violation detecting system according to Embodiment 3 of the present invention.
  • DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS In order to make the technical problems, technical solutions, and advantages of the present invention more comprehensible, the following detailed description will be made in conjunction with the accompanying drawings and specific embodiments.
  • the embodiment of the present invention is directed to the prior art, where the video recording and recording recording needs to be viewed by the user to record whether the violation behavior is recorded, and a multi-target reverse violation detection method, device and system are provided, which can be based on the monitoring direction set by the user.
  • FIG. 1 is a schematic flowchart of a multi-target reverse violation detection method according to Embodiment 1 of the present invention.
  • the detection method includes: Step S100: acquiring a video image to be detected; Step S102, receiving a user operation instruction, where Determining, in the detected video image, a region to be detected and a reference motion direction corresponding to each region to be detected; Step S104, determining a dynamic target in the to-be-detected area for each to-be-detected area; Step S106, determining, for each dynamic target, a moving direction of the dynamic target within a preset time period; Step S108, for each And a dynamic target, determining, according to the reference motion direction corresponding to the to-be-detected area to which the dynamic target belongs, whether the moving direction of the dynamic target in the preset time period is violated.
  • the multi-objective reverse violation detection method provided by the present invention can determine the to-be-detected area and the reference motion direction corresponding to each to-be-detected area in the video image to be detected according to the user operation instruction, according to the motion of the dynamic target in the area to be detected.
  • the direction and the reference motion direction corresponding to the to-be-detected area are compared and analyzed, and it is determined whether there is a reverse violation behavior in the to-be-detected area, thereby eliminating the need for the user to view whether there is a reverse violation behavior in the video image, thereby saving manpower.
  • the multi-target reverse violation detection method wherein the determining the dynamic target in the to-be-detected area may be specifically: acquiring the foreground data in the to-be-detected area by using an inter-frame difference multiplication method; The open/close operation processes the foreground data; and performs the connected region on the processed foreground data to obtain a dynamic target in the to-be-detected region.
  • the foreground data in the area to be detected which is obtained by the method of inter-frame difference multiplication, may be obtained by using the inter-frame difference multiplication method.
  • the foreground data can be processed by the open/close operation in the image morphology to eliminate the "cavity phenomenon".
  • the foreground data processed in this way may lead to some unconnected component merging. Therefore, it is necessary to use the connected region segmentation technique to split the merged components to obtain each dynamic target. Since the amount of data in the video image is too large, if the video image is directly analyzed, the performance is affected. Therefore, the multi-target reverse violation detection method, wherein the video image to be detected is binarized Value image.
  • the multi-target reverse violation detection method wherein the determining the moving direction of the dynamic target in the preset time period may be specifically: each frame image in the preset time period for the to-be-detected video image Determining the dynamic target position; determining a direction of motion of the dynamic target within the preset time period by direction fitting according to a position of the dynamic target in each frame image.
  • the purpose of the above operation is to perform motion direction detection for each dynamic target. Since the moving direction of the dynamic target is not obvious in a short time, the moving direction of the dynamic target in a certain period of time is usually judged.
  • the video image corresponds to the multi-frame image in the time period.
  • the multi-objective reverse violation detection method wherein the determination Whether the moving direction of the dynamic target in the preset time period is irregular may be specifically: determining an angle between the moving direction and the reference moving direction; when the angle exceeds a preset angle range, determining The dynamic target reverse violation.
  • the detection method includes: Step S200: Acquire a task to be detected, and acquire a to-be-detected video according to the information to be detected.
  • Step S202 determining a to-be-detected area and a reference moving direction on the to-be-detected video;
  • Step S204 reading one frame of the image to be detected;
  • Step S206 extracting foreground data in the image, and performing graphic form on the foreground data
  • Step S208 the location information of the dynamic target is recorded;
  • Step S210 it is determined whether the location information of the dynamic target exceeds 20, when it is exceeded, the process proceeds to step S212, otherwise, the process proceeds to step S204;
  • S212 fitting a trajectory of the dynamic target to obtain a moving direction of the dynamic target; and
  • step S214 calculating an angle between the moving direction and the reference moving direction;
  • step S216 it is determined whether the dynamic target is reversed.
  • the present invention also provides a multi-target reverse violation detection apparatus, comprising: an acquisition module, configured to acquire a video image to be detected; a first determining module, configured to receive a user operation instruction, in the to-be-detected video image, determining to be a detection area and a reference movement direction corresponding to each to-be-detected area; a second determining module, configured to determine a dynamic target in the to-be-detected area for each to-be-detected area; and a third determining module, configured to a dynamic target, determining a moving direction of the dynamic target in a preset time period; and a determining module, configured to determine, according to a reference moving direction corresponding to the to-be-detected area to which the dynamic target belongs, for each dynamic target Whether the target's direction of motion within the preset
  • the multi-target reverse violation detecting apparatus wherein the second determining module includes: an acquiring unit configured to acquire foreground data in the to-be-detected area by using an interframe difference multiplication method; and a preprocessing unit configured to pass the image form The open/close operation in the process processes the foreground data; the area dividing unit is configured to perform the connected area segmentation on the processed foreground data to obtain a dynamic target in the area to be detected.
  • the multi-target reverse violation detecting apparatus wherein the video image to be detected is a binary image subjected to binarization processing.
  • the multi-target reverse violation detecting device further comprising: a photographing module, configured to capture a snapshot of the dynamic target when the dynamic target violates the rule when the dynamic target violates a movement direction within a preset time period And store it.
  • the present invention also provides a multi-target reverse violation detection system, comprising the multi-target reverse violation detection device and a plurality of cameras, wherein the plurality of cameras are configured to generate a video image, and according to the requirements of the multi-target reverse violation detection device Take a snapshot capture.
  • 3 is a schematic structural diagram of a multi-target reverse violation detection system according to Embodiment 3 of the present invention.
  • the multi-target reverse violation detection system 30 includes: a video monitoring module 31: configured to generate a video image and perform snapshot capture.
  • the task access module 32 is configured to receive the detection task information set by the user, and according to the detection task information and the load capacity of the algorithm main control module, split the detection task information, generate multiple detection tasks, and generate a plurality of detection tasks from the video monitoring module.
  • the detection task may be obtained from an XML file configured by a user; the algorithm main control module 33 is configured to manage multiple algorithm modules, and the plurality of detection tasks split by the task access module 32 are The load capacity of each of the managed algorithm modules is sent to each algorithm module; the algorithm module 34: is configured to implement the function of the multi-target reverse detection device described above, and performs a detection task; and the database 35: is configured to store snapshot snapshots and violation information.
  • a multi-target reverse violation detection method, apparatus, and system provided by an embodiment of the present invention have the following beneficial effects: It is possible to automatically determine whether there is a reverse violation behavior in a video recording according to a monitoring direction set by a user. Therefore, there is no need for the user to check whether there is a reverse violation in the video recording, which saves manpower.

Abstract

一种多目标逆向违规检测方法、装置及系统,所述多目标逆向违规检测方法包括:获取待检测视频图像;接收用户操作指令,在所述待检测视频图像中,确定待检测区域及与每一待检测区域对应的基准运动方向;针对每一待检测区域,确定所述待检测区域内的动态目标;针对每一动态目标,确定所述动态目标在预设时间段内的运动方向;针对每一动态目标,根据与所述动态目标所属的待检测区域对应的基准运动方向,判断所述动态目标在预设时间段内的运动方向是否违规。能够根据用户设置的监测方向,自动判断视频录像中是否存在逆向违规的行为,从而,无需由用户查看视频录像中是否存在逆向违规的行为,节省了人力。

Description

一种多目标逆向违规检测方法、 装置及系统 技术领域 本发明涉及视频识别领域,尤其涉及一种多目标逆向违规检测方法、装置及系统。 背景技术 目前, 视频监控被用于各种场合中, 如商场、 地铁、 客运站出入口、 交通要道等, 对各种场合发生的事件进行监控, 然而, 在现有技术中, 主要提供的是一种事后的, 录像回退的方式来处理一些突发事件, 查看视频录像中是否存在违规的行为, 这样不 仅浪费人力, 耗损时间, 也会遗漏掉一部分有违规行为的路人或者车辆。 发明内容 本发明实施例提供了一种多目标逆向违规检测方法、 装置及系统, 以至少解决上 述技术问题, 能够根据用户设置的监测方向, 自动判断视频录像中是否存在逆向违规 的行为, 从而, 无需由用户查看视频录像中是否存在逆向违规的行为, 节省了人力。 为了实现上述目的, 本发明实施例的一个方面提供了一种多目标逆向违规检测方 法, 包括: 获取待检测视频图像; 接收用户操作指令, 在所述待检测视频图像中, 确 定待检测区域及与每一待检测区域对应的基准运动方向; 针对每一待检测区域, 确定 所述待检测区域内的动态目标; 针对每一动态目标, 确定所述动态目标在预设时间段 内的运动方向; 针对每一动态目标, 根据与所述动态目标所属的待检测区域对应的基 准运动方向, 判断所述动态目标在预设时间段内的运动方向是否违规。 上述多目标逆向违规检测方法, 其中, 所述确定所述待检测区域内的动态目标具 体为: 通过帧间差乘方法, 获取所述待检测区域内的前景数据; 通过图像形态学中的 开 /闭运算对所述前景数据进行处理; 对经过处理后的前景数据进行连通区域分割, 得 到所述待检测区域内的动态目标。 上述多目标逆向违规检测方法, 其中, 所述待检测视频图像为经过二值化处理的 二值图像。 上述多目标逆向违规检测方法, 其中, 所述确定所述动态目标在预设时间段内的 运动方向具体为: 针对所述待检测视频图像在所述预设时间段内的每一帧图像, 确定 所述动态目标位置; 根据所述动态目标在所述每一帧图像中的位置, 通过方向拟合确 定所述动态目标在所述预设时间段内的运动方向。 上述多目标逆向违规检测方法, 其中, 所述判断所述动态目标在预设时间段内的 运动方向是否违规具体为: 确定所述运动方向与所述基准运动方向之间的夹角; 当所 述夹角超出预设角度范围时, 判定所述动态目标逆向违规。 上述多目标逆向违规检测方法, 其中, 所述判断所述动态目标在预设时间段内的 运动方向是否违规之后还包括: 当所述动态目标在预设时间段内的运动方向违规时, 在所述动态目标违规时, 抓拍所述动态目标的快照并存储。 本发明实施例的另一个方面还提供了一种多目标逆向违规检测装置, 包括: 获取 模块, 设置为获取待检测视频图像; 第一确定模块, 设置为接收用户操作指令, 在所 述待检测视频图像中, 确定待检测区域及与每一待检测区域对应的基准运动方向; 第 二确定模块, 设置为针对每一待检测区域, 确定所述待检测区域内的动态目标; 第三 确定模块, 设置为针对每一动态目标,确定所述动态目标在预设时间段内的运动方向; 判断模块, 设置为针对每一动态目标, 根据与所述动态目标所属的待检测区域对应的 基准运动方向, 判断所述动态目标在预设时间段内的运动方向是否违规。 上述多目标逆向违规检测装置, 其中, 所述第二确定模块包括: 获取单元, 设置 为通过帧间差乘方法, 获取所述待检测区域内的前景数据; 预处理单元, 设置为通过 图像形态学中的开 /闭运算对所述前景数据进行处理; 区域分割单元, 设置为对经过处 理后的前景数据进行连通区域分割, 得到所述待检测区域内的动态目标。 上述多目标逆向违规检测装置, 其中, 所述待检测视频图像为经过二值化处理的 二值图像。 上述多目标逆向违规检测装置, 其中, 所述第三确定模块包括: 第一确定单元, 设置为针对所述待检测视频图像在所述预设时间段内的每一帧图像, 确定所述动态目 标位置; 第二确定单元, 设置为根据所述动态目标在所述每一帧图像中的位置, 通过 方向拟合确定所述动态目标在所述预设时间段内的运动方向。 上述多目标逆向违规检测装置, 其中, 所述判断模块包括: 第三确定单元, 设置 为确定所述运动方向与所述基准运动方向之间的夹角; 判定单元, 设置为当所述夹角 超出预设角度范围时, 判定所述动态目标逆向违规。 上述多目标逆向违规检测装置, 其中, 还包括: 拍照模块, 设置为当所述动态目 标在预设时间段内的运动方向违规时, 在所述动态目标违规时, 抓拍所述动态目标的 快照并存储。 本发明实施例的另一个方面还提供了一种多目标逆向违规检测系统, 包括上述多 目标逆向违规检测装置及多个摄像头, 所述多个摄像头设置为生成视频图像, 并根据 所述多目标逆向违规检测装置的需求进行快照抓拍。 本发明实施例的上述技术方案的有益效果如下: 本发明实施例提供了一种多目标逆向违规检测方法、 装置及系统, 能够根据用户 设置的监测方向, 自动判断视频录像中是否存在逆向违规的行为, 从而, 无需由用户 查看视频录像中是否存在逆向违规的行为, 节省了人力。 附图说明 图 1为本发明实施例 1提供的多目标逆向违规检测方法的流程示意图。 图 2为本发明实施例 2提供的多目标逆向违规检测方法的流程示意图。 图 3为本发明实施例 3提供的多目标逆向违规检测系统的结构示意图。 具体实施方式 为使本发明要解决的技术问题、 技术方案和优点更加清楚, 下面将结合附图及具 体实施例进行详细描述。 本发明实施例针对现有技术中, 对视频监控记录的录像需要由用户查看是否记录 了违规行为的问题, 提供了一种多目标逆向违规检测方法、 装置及系统, 能够根据用 户设置的监测方向, 自动判断视频录像中是否存在逆向违规的行为, 从而, 无需由用 户查看视频录像中是否存在逆向违规的行为, 节省了人力。 图 1为本发明实施例 1提供的多目标逆向违规检测方法的流程示意图,如图所示, 所述检测方法包括: 步骤 S100, 获取待检测视频图像; 步骤 S102, 接收用户操作指令, 在所述待检测视频图像中, 确定待检测区域及与 每一待检测区域对应的基准运动方向; 步骤 S104, 针对每一待检测区域, 确定所述待检测区域内的动态目标; 步骤 S106, 针对每一动态目标, 确定所述动态目标在预设时间段内的运动方向; 步骤 S108, 针对每一动态目标, 根据与所述动态目标所属的待检测区域对应的基 准运动方向, 判断所述动态目标在预设时间段内的运动方向是否违规。 本发明提供的多目标逆向违规检测方法, 能够根据用户操作指令, 在待检测视频 图像中确定待检测区域及与每一待检测区域对应的基准运动方向, 根据待检测区域内 的动态目标的运动方向及与该待检测区域对应的基准运动方向进行对比分析, 判断在 该待检测区域内是否存在逆向违规的行为, 从而无需由用户查看视频图像中是否存在 逆向违规的行为, 节省了人力。 上述多目标逆向违规检测方法, 其中, 所述确定所述待检测区域内的动态目标可 以具体为: 通过帧间差乘方法, 获取所述待检测区域内的前景数据; 通过图像形态学 中的开 /闭运算对所述前景数据进行处理; 对经过处理后的前景数据进行连通区域分 害 得到所述待检测区域内的动态目标。 为了得到待检测区域内的动态目标, 需要对待检测区域进行前景提取, 可以通过 帧间差乘的方法, 首先获取的待检测区域内的前景数据, 由于采用帧间差乘的方法所 获取的前景数据中往往会出现"空洞现象", 因此, 需要对该前景数据进行预处理, 消 除 "空洞现象", 可以采用图像形态学中的开 /闭运算对前景数据进行处理以消除 "空 洞现象", 在经过这样处理后的前景数据, 可能会导致一些不连通的分量合并, 因此, 需要采用连通区域分割技术, 将合并的分量拆分, 得到各个动态目标。 由于视频图像中的数据量过大, 如果对视频图像直接进行分析,则会影响到性能, 因此, 上述多目标逆向违规检测方法, 其中, 所述待检测视频图像为经过二值化处理 的二值图像。 上述多目标逆向违规检测方法, 其中, 所述确定所述动态目标在预设时间段内的 运动方向可以具体为: 针对所述待检测视频图像在所述预设时间段内的每一帧图像, 确定所述动态目标位置; 根据所述动态目标在所述每一帧图像中的位置, 通过方向拟 合确定所述动态目标在所述预设时间段内的运动方向。 上述操作的目的是, 对每个动态目标进行运动方向检测, 由于在短时间内对动态 目标的运动方向不明显, 因此, 通常会判断动态目标在一定的时间段内的运动方向, 视频图像在该时间段内对应多帧图像, 通过获取到动态目标在每一帧图像中的位置, 运用向量三角法对该多帧图像中的位置信息进行拟合, 得到该动态目标的运动方向。 由于在判断违规行为时, 通常只需要对动态目标的运动的趋势方向进行判断, 无 需要求动态目标的运动方向与基准运动方向完全一致, 因此, 上述多目标逆向违规检 测方法, 其中, 所述判断所述动态目标在预设时间段内的运动方向是否违规可以具体 为: 确定所述运动方向与所述基准运动方向之间的夹角; 当所述夹角超出预设角度范 围时, 判定所述动态目标逆向违规。 上述操作的目的是, 通过根据动态目标的运动方向与基准运动方向之间的夹角, 确定动态目标是否逆向违规。 上述多目标逆向违规检测方法, 其中, 所述判断所述动态目标在预设时间段内的 运动方向是否违规之后还可以包括: 当所述动态目标在预设时间段内的运动方向违规 时, 在所述动态目标违规时, 抓拍所述动态目标的快照并存储。 上述操作的目的是, 将违规的动态目标的违规行为进行拍照保存, 以作为证据进 行后续处理, 避免了需要证据时, 再次去查找视频录像。 图 2为本发明实施例 2提供的多目标逆向违规检测方法的流程示意图,如图所示, 所述检测方法包括: 步骤 S200, 获取到待检测任务信息, 根据待检测任务信息获取待检测视频; 步骤 S202, 在待检测视频上, 确定待检测区域及基准运动方向; 步骤 S204, 读取待检测视频的一帧图像; 步骤 S206,在所述图像中提取前景数据,对前景数据进行图形形态学的开 /闭操作 得当前景动态目标; 步骤 S208, 记录动态目标的位置信息; 步骤 S210, 判断动态目标的位置信息是否超出 20个, 当超出时, 进入步骤 S212, 否则, 进入步骤 S204; 步骤 S212, 对动态目标的轨迹进行拟合, 得到动态目标的运动方向; 步骤 S214, 计算运动方向与基准运动方向的夹角; 步骤 S216, 判断动态目标是否逆向违规, 当违规时, 进入步骤 S218, 否则, 进 入步骤 204; 步骤 S218, 上报违规告警信息, 并对动态目标的违规行为进行快照抓拍。 本发明还提供了一种多目标逆向违规检测装置, 包括: 获取模块, 设置为获取待 检测视频图像; 第一确定模块, 设置为接收用户操作指令, 在所述待检测视频图像中, 确定待检测区域及与每一待检测区域对应的基准运动方向; 第二确定模块, 设置为针 对每一待检测区域, 确定所述待检测区域内的动态目标; 第三确定模块, 设置为针对 每一动态目标, 确定所述动态目标在预设时间段内的运动方向; 判断模块, 设置为针 对每一动态目标, 根据与所述动态目标所属的待检测区域对应的基准运动方向, 判断 所述动态目标在预设时间段内的运动方向是否违规。 上述多目标逆向违规检测装置, 其中, 所述第二确定模块包括: 获取单元, 设置 为通过帧间差乘方法, 获取所述待检测区域内的前景数据; 预处理单元, 设置为通过 图像形态学中的开 /闭运算对所述前景数据进行处理; 区域分割单元, 设置为对经过处 理后的前景数据进行连通区域分割, 得到所述待检测区域内的动态目标。 上述多目标逆向违规检测装置, 其中, 所述待检测视频图像为经过二值化处理的 二值图像。 上述多目标逆向违规检测装置, 其中, 所述第三确定模块包括: 第一确定单元, 设置为针对所述待检测视频图像在所述预设时间段内的每一帧图像, 确定所述动态目 标位置; 第二确定单元, 设置为根据所述动态目标在所述每一帧图像中的位置, 通过 方向拟合确定所述动态目标在所述预设时间段内的运动方向。 上述多目标逆向违规检测装置, 其中, 所述判断模块包括: 第三确定单元, 设置 为确定所述运动方向与所述基准运动方向之间的夹角; 判定单元, 设置为当所述夹角 超出预设角度范围时, 判定所述动态目标逆向违规。 上述多目标逆向违规检测装置, 其中, 还包括: 拍照模块, 设置为当所述动态目 标在预设时间段内的运动方向违规时, 在所述动态目标违规时, 抓拍所述动态目标的 快照并存储。 本发明还提供了一种多目标逆向违规检测系统, 包括上述多目标逆向违规检测装 置及多个摄像头, 所述多个摄像头设置为生成视频图像, 并根据所述多目标逆向违规 检测装置的需求进行快照抓拍。 图 3为本发明实施例 3提供的多目标逆向违规检测系统的结构示意图,如图所示, 所述多目标逆向违规检测系统 30包括: 视频监控模块 31 : 设置为生成视频图像并进行快照抓拍; 任务接入模块 32: 设置为接收用户设置的检测任务信息, 并根据该检测任务信息 及算法主控模块的负载能力, 对检测任务信息进行拆分, 生成多个检测任务并从视频 监控模块中获取对应的视频图像, 该检测任务可以从用户配置的 XML文件中获取; 算法主控模块 33 : 设置为管理多个算法模块, 并将任务接入模块 32拆分的多个 检测任务根据其所管理的各个算法模块的负载能力下发至各个算法模块; 算法模块 34: 设置为实现上述的多目标逆向检测装置的功能, 执行检测任务; 数据库 35 : 设置为存储抓拍的快照及违规信息。 以上所述是本发明的优选实施方式, 应当指出, 对于本技术领域的普通技术人员 来说, 在不脱离本发明所述原理的前提下, 还可以作出若干改进和润饰, 这些改进和 润饰也应视为本发明的保护范围。 工业实用性 如上所述, 本发明实施例提供的一种多目标逆向违规检测方法、 装置及系统, 具有以下有益效果: 能够根据用户设置的监测方向, 自动判断视频录像中是否存在 逆向违规的行为, 从而, 无需由用户查看视频录像中是否存在逆向违规的行为, 节 省了人力。

Claims

权 利 要 求 书
1 . 一种多目标逆向违规检测方法, 包括: 获取待检测视频图像;
接收用户操作指令, 在所述待检测视频图像中, 确定待检测区域及与每一 待检测区域对应的基准运动方向; 针对每一待检测区域, 确定所述待检测区域内的动态目标;
针对每一动态目标, 确定所述动态目标在预设时间段内的运动方向; 针对每一动态目标, 根据与所述动态目标所属的待检测区域对应的基准运 动方向, 判断所述动态目标在预设时间段内的运动方向是否违规。
2. 如权利要求 1所述方法,其中,所述确定所述待检测区域内的动态目标具体为: 通过帧间差乘方法, 获取所述待检测区域内的前景数据; 通过图像形态学中的开 /闭运算对所述前景数据进行处理;
对经过处理后的前景数据进行连通区域分割, 得到所述待检测区域内的动 态目标。
3. 如权利要求 1所述方法, 其中, 所述待检测视频图像为经过二值化处理的二值 图像。
4. 如权利要求 1所述方法, 其中, 所述确定所述动态目标在预设时间段内的运动 方向具体为:
针对所述待检测视频图像在所述预设时间段内的每一帧图像, 确定所述动 态目标位置;
根据所述动态目标在所述每一帧图像中的位置, 通过方向拟合确定所述动 态目标在所述预设时间段内的运动方向。
5. 如权利要求 1所述方法, 其中, 所述判断所述动态目标在预设时间段内的运动 方向是否违规具体为:
确定所述运动方向与所述基准运动方向之间的夹角;
当所述夹角超出预设角度范围时, 判定所述动态目标逆向违规。
6. 如权利要求 1所述方法, 其中, 所述判断所述动态目标在预设时间段内的运动 方向是否违规之后还包括: 当所述动态目标在预设时间段内的运动方向违规时, 在所述动态目标违规 时, 抓拍所述动态目标的快照并存储。
7. 一种多目标逆向违规检测装置, 包括: 获取模块, 设置为获取待检测视频图像;
第一确定模块, 设置为接收用户操作指令, 在所述待检测视频图像中, 确 定待检测区域及与每一待检测区域对应的基准运动方向;
第二确定模块, 设置为针对每一待检测区域, 确定所述待检测区域内的动 态目标;
第三确定模块, 设置为针对每一动态目标, 确定所述动态目标在预设时间 段内的运动方向; 判断模块, 设置为针对每一动态目标, 根据与所述动态目标所属的待检测 区域对应的基准运动方向, 判断所述动态目标在预设时间段内的运动方向是否 违规。
8. 如权利要求 7所述装置, 其中, 所述第二确定模块包括: 获取单元,设置为通过帧间差乘方法,获取所述待检测区域内的前景数据; 预处理单元,设置为通过图像形态学中的开 /闭运算对所述前景数据进行处 理; 区域分割单元, 设置为对经过处理后的前景数据进行连通区域分割, 得到 所述待检测区域内的动态目标。
9. 如权利要求 7所述装置, 其中, 所述待检测视频图像为经过二值化处理的二值 图像。
10. 如权利要求 7所述装置, 其中, 所述第三确定模块包括: 第一确定单元, 设置为针对所述待检测视频图像在所述预设时间段内的每 一帧图像, 确定所述动态目标位置; 第二确定单元, 设置为根据所述动态目标在所述每一帧图像中的位置, 通 过方向拟合确定所述动态目标在所述预设时间段内的运动方向。
11 . 如权利要求 7所述装置, 其中, 所述判断模块包括: 第三确定单元,设置为确定所述运动方向与所述基准运动方向之间的夹角; 判定单元, 设置为当所述夹角超出预设角度范围时, 判定所述动态目标逆 向违规。
12. 如权利要求 7所述装置, 其中, 还包括: 拍照模块, 设置为当所述动态目标在预设时间段内的运动方向违规时, 在 所述动态目标违规时, 抓拍所述动态目标的快照并存储。
13. 一种多目标逆向违规检测系统,包括如权利要求 7-12任一项所述的多目标逆向 违规检测装置及多个摄像头, 所述多个摄像头设置为生成视频图像, 并根据所 述多目标逆向违规检测装置的需求进行快照抓拍。
PCT/CN2014/082767 2014-05-26 2014-07-22 一种多目标逆向违规检测方法、装置及系统 WO2015180252A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201410226131.1A CN105224905A (zh) 2014-05-26 2014-05-26 一种多目标逆向违规检测方法、装置及系统
CN201410226131.1 2014-05-26

Publications (1)

Publication Number Publication Date
WO2015180252A1 true WO2015180252A1 (zh) 2015-12-03

Family

ID=54697970

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2014/082767 WO2015180252A1 (zh) 2014-05-26 2014-07-22 一种多目标逆向违规检测方法、装置及系统

Country Status (2)

Country Link
CN (1) CN105224905A (zh)
WO (1) WO2015180252A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107786837B (zh) * 2016-08-24 2021-06-04 杭州海康威视数字技术股份有限公司 扶梯危险状态检测方法及装置
CN107194360B (zh) * 2017-05-25 2018-07-20 智慧航安(北京)科技有限公司 逆向通行对象识别方法、装置及系统
CN108491795B (zh) * 2018-03-22 2022-05-13 北京航空航天大学 轨道交通场景的行人检测方法与装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2003028376A1 (en) * 2001-09-14 2003-04-03 Vislog Technology Pte Ltd Customer service counter/checkpoint registration system with video/image capturing, indexing, retrieving and black list matching function
CN101093603A (zh) * 2007-07-03 2007-12-26 北京智安邦科技有限公司 智能视频监控设备模组和系统及其监控方法
CN101159859A (zh) * 2007-11-29 2008-04-09 北京中星微电子有限公司 一种运动检测方法、装置以及一种智能监控系统
CN101465033A (zh) * 2008-05-28 2009-06-24 丁国锋 一种自动追踪识别系统及方法
CN101763388A (zh) * 2008-12-25 2010-06-30 北京中星微电子有限公司 视频检索方法、系统、设备及视频存储方法、系统

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101308576B (zh) * 2008-07-11 2011-04-06 北京中星微电子有限公司 道路中的运动方向检测方法及装置
US9024785B2 (en) * 2011-12-27 2015-05-05 Aisin Aw Co., Ltd. Traffic information distribution system and traffic information system, traffic information distribution program, and traffic information distribution method
CN103414870B (zh) * 2013-07-16 2016-05-04 南京师范大学 一种多模式警戒分析方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2003028376A1 (en) * 2001-09-14 2003-04-03 Vislog Technology Pte Ltd Customer service counter/checkpoint registration system with video/image capturing, indexing, retrieving and black list matching function
CN101093603A (zh) * 2007-07-03 2007-12-26 北京智安邦科技有限公司 智能视频监控设备模组和系统及其监控方法
CN101159859A (zh) * 2007-11-29 2008-04-09 北京中星微电子有限公司 一种运动检测方法、装置以及一种智能监控系统
CN101465033A (zh) * 2008-05-28 2009-06-24 丁国锋 一种自动追踪识别系统及方法
CN101763388A (zh) * 2008-12-25 2010-06-30 北京中星微电子有限公司 视频检索方法、系统、设备及视频存储方法、系统

Also Published As

Publication number Publication date
CN105224905A (zh) 2016-01-06

Similar Documents

Publication Publication Date Title
US10212397B2 (en) Abandoned object detection apparatus and method and system
CN102665071B (zh) 一种社会治安视频监控图像智能处理和检索方法
Jadhav et al. Smart traffic control system using image processing
KR102217253B1 (ko) 행동패턴 분석 장치 및 방법
CN204155449U (zh) 一种车牌识别比对高清摄像机及监控系统
CN110796819B (zh) 一种针对站台黄线入侵越界人员的检测方法及系统
Zabłocki et al. Intelligent video surveillance systems for public spaces–a survey
CN111310645A (zh) 货物堆积量的溢仓预警方法、装置、设备及存储介质
CN105426869A (zh) 基于铁路安检的人脸识别系统及识别方法
CN111259868B (zh) 基于卷积神经网络的逆行车辆检测方法、系统及介质
Kulkarni et al. Real time vehicle detection, tracking and counting using Raspberry-Pi
KR101492473B1 (ko) 사용자 기반 상황 인지형 씨씨티비 통합관제시스템
Doshi et al. An efficient approach for anomaly detection in traffic videos
WO2015180252A1 (zh) 一种多目标逆向违规检测方法、装置及系统
CN109643488A (zh) 交通异常事件检测装置及方法
Im et al. Bigdata analytics on CCTV images for collecting traffic information
CN105227918A (zh) 一种智能监控方法及装置
KR102434154B1 (ko) 영상감시시스템에서의 고속 이동물체의 위치 및 모션 캡쳐 방법
CN210666820U (zh) 一种基于dsp边缘计算的行人异常行为检测系统
TWI617997B (zh) Intelligent object detection assistance system and method
UrRehman et al. Modeling, design and analysis of intelligent traffic control system based on integrated statistical image processing techniques
Sofwan et al. Design of smart open parking using background subtraction in the IoT architecture
CN112686136A (zh) 一种对象检测方法、装置及系统
WO2012074366A2 (en) A system and a method for detecting a loitering event
KR20110102321A (ko) 이미지 움직임 검출 방법 및 장치

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 14893507

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 14893507

Country of ref document: EP

Kind code of ref document: A1